There currently are no established biomarkers that predict who among asymptomatic individuals will develop Alzheimer's disease (AD), or who with memory impairment will convert to AD. Finding such a biomarker is critical in identifying individuals for treatment early in the disease course;once symptoms begin, the pathology may be irreversible. AD is a progressive neurodegenerative disorder and membrane lipids and lipid metabolites are involved in the neurodegeneration. Animal and human laboratory studies suggest associations between ceramides, sphingomyelins, 24S-hydroxycholesterol (24S-OHC), or F2a-Isoprostanes and neurodegeneration. The few available clinic studies, and our preliminary data, suggest these lipids are altered in blood, CSF and brain in the earliest stages of AD (CDR 0.5), and vary by AD severity. We hypothesize these lipids are markers of ongoing neurodegeneration, and preclinical risk factors or indicators of impending AD;thereby holding promise as candidate biomarkers of AD progression. If they could be developed into blood-based biomarkers, they would be superior to CSF or neuroimaging measures with regards to cost, invasiveness, and feasibility for repeated measures. The overall aim of the proposed study is to extend this line of laboratory and clinical research to a well-characterized, longitudinal, population-based study. Most AD biomarker studies have focused on diagnostic markers, incorporating cross-sectional studies of clinic populations. While clinic studies are helpful in determining whether a biomarker varies by AD severity, a longitudinal population-based study is critical to discern whether a biomarker may best be suited to the earlier stages of the disease, whether it is predictive of the rate of progression, and to provide descriptive data on the variability of the biomarker. To this regard, we will examine the above-mentioned serum lipids in data already collected from older women enrolled in the Women's Health and Aging Study (WHAS) II. This population- based longitudinal study will allow us to examine whether baseline measures of these lipids predict subsequent cognitive impairment across domains and incident dementia (AD and all-cause) over 9 years of follow-up, thereby identifying a potential pre-clinical biomarker. The PI previously obtained a pilot grant to develop the lipid assays using baseline serum from WHAS II. We now apply for funds to conduct the secondary data analysis. Specifically, analyses will: (1) Characterize the between-individual variability of the serum lipids and factors that affect this variability such as demographic factors, medical conditions, smoking, and APOE e4;(2) Examine the cross-sectional and longitudinal associations between these lipids and incident impairment on tests of specific cognitive domains;and (3) Examine the relationship between the serum lipids and incident AD. PUBLIC HEALTH RELEVANCE: This R03 application proposes to extend the current line of laboratory and clinical research examining membrane lipids and metabolites (ceramides, sphingomyelins, 24S-hydroxycholesterol (24S-OHC), and F2a-Isoprostanes) to a longitudinal, population-based study in order to determine whether they are predictive of subsequent cognitive impairment and Alzheimer's disease (AD). Most AD biomarker studies have focused on diagnostic markers, incorporating cross-sectional studies of clinic populations. While clinic studies are helpful in determining whether a biomarker varies by AD severity, a longitudinal population-based study is critical to discern whether a biomarker may best be suited to the earlier stages of the disease, whether it is predictive of the rate of progression, and to provide descriptive data on the variability of the biomarker.